INDEX

Explanations

references to unfamiliar individuals or characters

oai_token-act-pair · gpt-4o-mini Triggered by @bot

people not known before

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Ambiental

-0.40

">//

-0.38

AddTagHelper

-0.37

 كومونز

-0.36

 onCancelled

-0.34

 ejecutivo

-0.34

 deberes

-0.33

 mittler

-0.33

RequiresApi

-0.33

 gobernador

-0.32

POSITIVE LOGITS

 strangers

0.85

 stranger

0.75

ArgsConstructor

0.75

 Strangers

0.72

stranger

0.69

陌生

0.66

 unfamiliar

0.65

 acquaintances

0.63

 ſind

0.63

 acquaintance

0.60

Activations Density 0.023%

references to unfamiliar individuals or characters

people not known before

No Comments

No Known Activations